VoxEU Column Frontiers of economic research Labour Markets

On work hours in the US and Europe

It is no secret that Americans work more than Europeans – 30% more according to recent studies. Many economists point to higher taxes in Europe as a major cause. This column suggests that divorce rates also play a role, particularly for women's labour supply.

According to recent research, Americans work 30% more than Europeans (Prescott 2004 and Rogerson 2006). This was not the case in early 1970s when Western Europeans worked more than Americans. What accounts for the large differences between countries today? Our study finds that divorce rates and tax rates together on average explain 58% of the difference in terms of hours worked between the US and 17 European countries.

Our research starts by trying to uncover the determinants of cross-country differences in work hours through analysing the hours worked by different demographic subgroups (Chakraborty et al. 2012). We find that women are typically the largest contributors to the aggregate differences. European women work less than American women, irrespective of whether we look at single or married women, or women with and without children.

Cross-country variation in tax systems is one of the most popular proposed explanations for the observed discrepancy in work hours between the US and Europe. The basic intuition here is straightforward – higher labour income taxes reduce the incentives to work. However, for the 18 countries in our sample, we find that, while there is indeed a negative correlation between various measures of taxation and hours worked by men, the corresponding correlation for women is close to 0. Figure 1 below plots the correlation between male and female labour supply and the "average effective tax rate", a measure that combines the average labour income tax and consumption tax in each country into a single tax rate.

Figure 1. Tax rates and labour supply by gender

At the same time, Figure 2 shows that we get the opposite result when looking at the relationship between divorce rates and work hours – we find a strong positive correlation between divorce rates and hours worked by women, whereas male hours and divorce rates are completely unrelated.

Figure 2. Divorce rate and labour supply by gender

While Figure 2 shows that female labour supply is correlated with divorce rates in a statistical sense, it does not answer the key question: Why would the likelihood of divorce affect the decision of females to work?

We believe this is because marriage provides an implicit social insurance since the spouses are able to share their income. However, if divorce rates are higher in a society, women have a higher incentive to obtain work experience in case they find themselves alone in the future. The reason the incentive is higher is because in our data, women happen to be the second earner in the household more often than men. European women anticipate not getting divorced as often and hence find less reason to insure themselves by working as much as American women.

We use this logic in a model that is developed to capture individual behaviour in terms of hours worked in response to anticipated divorce rates and tax rates. The individuals in our model start their working life at age 20. In every year, until retirement, they can decide how many hours to work, how much to save, and how much to consume. People who work accumulate work experience, which leads to higher earnings. We then simulate the lives of a cross section of individuals and study their behaviour as they are subjected to divorce/tax rates from the US and different European countries to obtain our main findings.

To evaluate the impact of the proposed channels, we calibrate our model to US data and study how labour supply in the US changes as we introduce divorce and marriage rates and tax systems of European countries. We find that stable marriages lead to a reduction in labour supply. Both in the data and in the model the effect is on the extensive margin, i.e. whether women will choose to work or not. In our counterfactual economy, if just marriage stability in US were the same as in the European countries, it explains 24% of the cross-country variation in hours worked for females. When we also introduce European taxes, we are able to explain 43% of the variation in female work hours across the continents.

Similarly to what we see in the data, we find that in our model taxes impact mostly men and along the intensive margin, i.e. the number of hours worked by employed people. The intuition behind this is that taxes are progressive, i.e. the more a person earns the higher is the marginal tax rate, and that the countries with the highest average tax rates also tend to have the most progressive taxes. This means that taxes on low earners (women) are often not so much higher than in the US while taxes on men are often much higher.

Our divorce argument is further supported by empirical studies finding that US states that adopted "no-fault" divorce laws in the 1970s experienced a spike in female labour supply relative to other states (no-fault divorces do not require a showing of wrong-doing by either party). This is also precisely the time when Americans gradually started to work more than the Europeans. Most European countries adopted no-fault divorce laws later and still have significantly lower divorce rates. Our tax argument finds support in studies pointing out that European and American tax levels were about equal around 1970 when they began to diverge.

We conclude that in response to changing social norms, American women have moved to insure their own future by working harder, while European men in response to higher taxes have found it less attractive to work long hours.

References

Chakraborty, Indraneel, Hans A Holter and Serhiy Stepanchuk (2012). “Marriage Stability, Taxation and Aggregate Labor Supply in the US vs. Europe”, Working Paper.

Prescott, Edward C (2004), “Why do Americans Work so Much More Than Europeans?”, Federal Reserve Bank of Minneapolis Quarterly Review, 28:2-13.

Rogerson, Richard (2006), “Understanding Differences in Hours Worked”, Review of Economic Dynamics, 365-409.

9,239 Reads